3 ms·
https://news.ycombinator.com/item?id=35333939 https://news.ycombinator.com/item?id=35333939 Have a read of this write up, it compliments this discussion nicely
by ChatGTP 4y ago
https://news.ycombinator.com/item?id=35333939 https://news.ycombinator.com/item?id=35333939
Have a read of this write up, it compliments this discussion nicely.
I think it makes some good points, but one of them stands out clearly to me, there is a narrative going around that we "have" to continue doing AI research at this pace. I'm skeptical about it. This is a story we believe.
We're fueling an arms race by doing the research, do you know what I mean? Like if we were throwing billions of dollars into gray goo creating technologies, would that be clever?
I'm not saying that it's possible to slow down or stop AI progress, but we're definitely not helping by throwing billions or trillions of dollars into it.
If we were an actually intelligent species, I think we would slow down and take stock of what we're doing before disaster strikes.
In fact global totalitarian surveillance that stamps out all potentially dangerous research is one of his few proposed solutions in his paper.
Someone else posted this, I actually think this is where further AI research will take us, this is because regular people who aren't nerds will vote for this, it would be politically an easy sell for a politician, and people would honestly prefer oversight and their families safety rather than having watching "boffins gone wild" receive endless rounds of funding creating self-replicating terminators for fun.
You have to remember, computer geeks and regular people probably have very different ideas about AI. If we can't behave, people will behave us.